differentiation-induced increase in clostridium botulinum c3 exoenzyme-catalyzed adp-ribosylation of...

Eur. J. Biochem. 223, 909-916 (1994) 0 FEBS 1994

Differentiation-induced increase in Clostridium botulinum C3 exoenzyme-catalyzed ADP-ribosylation of the small GTP-binding protein Rho Gerhard FRITZ, Ingo JUST, Peter WOLLENBERG and Klaus AKTORIES Institut fur Pharmakologie und Toxikologie der Universitat des Saarlandes, Homburg, Germany

(Received March 30May 17, 1994) - EJB 94 0442/1

The specific [‘2P]ADP-ribo~ylation by Clostridium botulinum exoenzyme C3 was used to study differentiation-dependent changes in the regulation of the low-molecular-mass GTP-binding protein Rho. Differentiation of F9 teratocarcinoma cells to neuronal-like cells by treatment with retinoic acid and dibutyryl-adenosine 3’,5’-monophosphate [(B t),cAMP] increased the C3-catalyzed ADP- ribosylation of RhoA proteins in cytosolic and membrane fractions by about threefold and sixfold, respectively. Phenotypical differentiation of F9 cells was not required for increase in ADP-ribosylation. Increase in ADP-ribosylation after (Bt),cAMP and retinoic acid treatments was blocked by cycloheximide, indicating the requirement of protein biosynthesis. As deduced from specific rho mRNA amounts and from Western analysis with a monoclonal RhoA antibody, the stimulation in the [32P]ADP-ribosylation of Rho was not caused by an increased de-novo synthesis of Rho proteins. GDP increased the ADP-ribosylation of membrane-associated Rho from non-differentiated, but not from differentiated F9 cells. GTP[S] decreased ADP-ribosylation of membranous Rho from differentiated and much less from non-differentiated F9 cells. Differentiation-dependent increase in ADP-ribosylation of cytosolic Rho was reversed by protein phosphatase type-1 . Treatment with SDS (0.01 %) which releases Rho from complexation with guanine nucleotide dissociation inhibitor, increased ADP-ribosylation both in differentiated and non-differentiated cells, indicating no differentiation-specific change of such complexes. In total, our data indicate that the induction of the differentiation process in F9 cells is accompanied by changes in the regulation of cytosolic and membrane-associated Rho proteins.

Various bacterial ADP-ribosyltransferases like Clostridium hotulinum C3 exoenzyme [l -31, C. limosum exoenzyme [4], Bacillus cereus exoenzyme [5] and a transferase from Staph- ylococcus aureus [6-81 have been shown to modify selec- tively the low-molecular-mass GTP-binding proteins RhoA, RhoB and RhoC [l-3, 9-17]. As known for other GTP- binding proteins, Rho proteins are regulated by a GTPase cycle. Various GTPase-activating proteins, guanine nucleotide dissociation-stimulating and guanine nucleotide dissociation-inhibiting proteins (GDI) have been identified [18 - 241 which are important for the conversion of Rho proteins from the GTP-bound (active) form to the GDP-bound (inactive) form and the reverse reaction. Mutation of Gly14 to Val ren- ders Rho unresponsive to GTPase-activating protein [20] and constitutively activates Rho protein. On the other hand, ADP-ribosylation of Rho by C3, which occurs in the putative

Correspondence to G. Fritz, Institut fur Pharmakologie und Toxikologie, Universitat des Saarlandes, D-66421 Homburg-Sax, Germany

Fax: +49 6841 16 6402. Abbreviations. (Bt),cAMP, dibutyry-adenosine 3’,5’-monophos-

phate; GDI, guanine nucleotide dissociation inhibitor; GST-RhoA, recombinant glutathione-S-transferase-RhoA fusion protein; GTP[S], guanosine 5’-[y-thioltriphosphate; PhMeSO,F, phenylmethylsulfonyl fluoride.

Enzymes. Glyceraldehyde-3-phosphate dehydrogenase (EC 1.2.1.12); glutathione S-transferase (EC 2.5.1.18).

effector region of Rho at position Asn41, results in the bio- logical inactivation of the GTP-binding protein 1251.

Rho proteins are apparently involved in the organization of the actin cytoskeleton. This was concluded from the findings that C3 exoenzyme caused loss of stress fibers and cell rounding, whereas microinjection of constitutively active [Vall4]Rho protein induces rapid reorganization of actin fila- ments [ 11,26,27]. Furthermore, Rho proteins have been pro- posed to be involved in cell motility [28, 291 and smooth muscle contraction [30]. Stable transfection of the complete rhoA cDNA into rat fibroblasts suggested that Rho proteins also participate in malignant transformation [3 1, 321.

So far, little is known about physiological conditions govering Rho activatiodinactivation or about the regulation cascade from Rho to the cytoskeleton. One approach to in- vestigate regulation of Rho on the protein level is the use of the clostridial exoenzyme C3. It is known that the extent of the ADP-ribosylation of Rho by C3 is influenced by the kind of guanine nucleotide bound [2, 5 , 14, 33, 341. Furthermore, the binding of the regulatory protein Rho-GDI reduces the ability of Rho to serve as substrate for C3 135, 361. Thus, measuring the ADP-ribosylation of Rho and analysing the molecular basis for the observed changes may be a useful tool to study physiological conditions which influence the regulation of Rho. Here we report that differentiation of F9 teratocarcinoma cells causes large changes in the ADP-ribo-

910

sylation of Rho, suggesting the involvement of Rho proteins in differentiation processes.

MATERIALS AND METHODS

Cell culture All cell lines used in the present study were obtained

from B. Kaina (Kernforschungszentrum Karlsruhe, Institut fur Genetik und Toxikologie, Germany). H4IIE and H5 rat hepatoma cells were cultivated in a-modified Eagle's medium containing 5 % heat-inactivated fetal calf serum. NIH3T3 (mouse fibroblasts) and F9 cells (mouse teratocarcinoma cells) were cultivated in Dulbecco's modified Eagle's medium containing 10% heat-inactivated fetal calf serum. Additionally, all media were supplemented with L-glutamine (2 mM), penicillin (100 unitdml) and streptomycin (100 pg/ ml).

Differentiation treatment H41IE and H5 rat hepatoma cells are described as cell

lines, which are distinguished in their differentiation state [37]. H4IIE shows a more differentiated, H5 a more dediffer- entiated phenotype. In contrast to these cells, in F9 teratocarcinoma cells the differentiation process can be induced by drug treatment with retinoic acid plus dibutyryl-adenosine 3',5'-monophosphate [(Bt)2cAMP] as described [38-401. Shortly, 24-48 h after seeding, the cells were treated with retinoic acid (0.1 pM) for 48 h. Subsequently, medium was replaced by new medium containing 2% fetal calf serum, retinoic acid (0.1 pM) and (Bt),cAMP (1 mM); 24-48 h afterwards, the phenotypic conversion to cells described as neuronal-like [38, 391 was observed.

Northern analysis Total RNA was prepared by lysis of the cells onto the

plates with guanidinium thiocyanate followed by acidphenol extraction [41]. Total RNA was electrophoretically separated on I % formamide gels and transferred overnight 20X NaCI/ Cit (175.3 g/l NaC1, 88.2 g/l sodium citrate pH 7.0) onto Hy- bond N' membrane (Amersham) as described [42]. RNA was fixed onto the membrane by incubation with 50mM NaOH for 5 min. Prehybridization was performed for 2 h in a solution containing 7% SDS (Bio-Rad), 1 mM EDTA, 0.5 M sodium phosphate pH 7.0. Hybridization was performed overnight in the same solution with additional 1% bovine serum albumin and with "P-labelled cDNA probe (specific activity about 10" cpm/p). The rhoA cDNA was kindly provided by A. Hall (Chester Beatty Laboratories, Institute of Cancer Research, London, UK), the glyceraldehyde-3-phosphate dehydrogenase and v$os cDNA probes by B. Kaina (Karlsruhe, Germany). After hybridization, the filters were washed once in 2X NaCUCit and twice in 1 X NaCUCit containing 0.5% SDS and 1 mM EDTA. All steps were performed at 65 "C. Filters were rehybridized with 32P-labelled glyceraldehyde-3-phosphate dehydrogenase probe as an internal standard.

ADP-ribosylation After harvesting, cells were disrupted by sonication in a

buffer containing 10 mM Tris/HCI pH 7.4, 1 mM EDTA, 1 mM MgCl,, 0.1 mM phenylmethylsulfonyl fluoride

(PhMeS0,F). After centrifugation (10 min, 600Xg, 4°C) su- pernatant was used for determination of the protein concentration according to Bradford [43]. ADP-ribosylation of Rho proteins by C3 was performed with 10-20 pg protein of the total cell extracts for 30 min at 37°C in a volume of 50 pl buffer A [20 mM Tris/HCI pH 7.4, 1 mM EDTA, 1 mM MgC12, 1 mM dithiothreitol, 10 mM thymidine, 0.2 pM [32P]NAD (0.5 pCi)] and 0.1 pg C3. C3 had been purified as described [1]. Reaction products were then analysed by SDS gel electrophoresis according to Laemmli [44] or by two- dimensional gel electrophoresis [45], followed by exposure of the dried gels to Kodak X-Omat films. To separate crude cytosolic fraction from crude membrane fraction, a further centrifugation step (20 min, 1 OOOOXg, 4°C) was performed.

Dephosphorylation by protein phosphatase type-1

Cytosolic extract (10 pg) was incubated with protein phosphatase type-1 (purified from bovine retina, provided by S. Klumpp, Tubingen, Germany) for 2 h at 37°C in a total volume of 10 pl buffer B (50 mM Tris/HCl pH 7.4, 1 mM EDTA, 20 mM MgC12, 0.1 mM PhMeSO,F, 0.5 mg/ml bovine serum albumin, 0.1 % 2-mercaptoethanol, 10 ng phosphatase/assay). As a control, extracts were incubated in the presence of dephosphorylation buffer without phosphatase. After dephosphorylation, ADP-ribosylation was performed as described before.

Nucleotide exchange reaction

Nucleotide exchange reactions were performed as described [14]. Shortly, protein extract was incubated in the presence of EDTA (5 mM) and 300 pM nucleotide (GDP or GTP[S]) for 10min at 30°C. Subsequently, MgCI, was added (final concentration 6 mM) and ADP-ribosylation performed as described above. Under these conditions, exchange efficiencies of > 90% were observed with the fusion protein of recombinant glutathione S-transferase with RhoA (GST- RhoA).

RhoA-binding analysis

By using the PCR technique, the complete RhoA coding sequence was amplified from rhoA plasmid 1201 (kindly provided by A. Hall, London, UK). The rhoA-specific oligonu- cleotides used for amplification contained additional BarnHI (5' primer) and EcoRI (3' primer) linker sequences. Using these restriction sites, the PCR product was cloned in the correct orientation into pGEX-2T expression plasmid (Phar- macia). The =50-kDa GST-RhoA fusion protein obtained was ADP-ribosylated by C3 and showed binding of GTP. The correct reading frame was also confirmed by sequencing. To screen for cellular proteins which might affect Rho ADP- ribosylation by binding to RhoA, 20-30 pg total cell extract was incubated with 200-300 ng recombinant GST-RhoA fusion protein for 30 min on ice. Subsequently, ADP-ribosylation by C3 was performed. The ADP-ribosylation of the fusion protein incubated with cell extract was compared with (a) the ADP-ribosylation of the fusion protein after preincu- bation with bovine serum albumin (instead of cell extract) and (b) the ADP-ribosylation of the endogenous Rho protein. The recombinant GST-RhoA fusion protein was used for these experiments because it is easy to distinguish from the endogenous Rho proteins by its higher molecular mass.

911

Western analysis After separation by SDS/PAGE, proteins were blotted on

nitrocellulose by semi-dry blotting. Subsequently, the filter was blocked with phosphate-buffered saline (NaCW, = 8 gA NaCI, 0.2 g/l KCI, 1.44 g/l Na,HPO,, 0.24 g KH,PO,/I pH 7.4), containing 1 % gelatin for 2 h at room temperature and then incubated overnight at 4°C with monoclonal Rho- GDI antibody JK-5 [46] (provided by Y. Takai, Kobe, Japan) or monoclonal Rho antibody [47] (provided by J. Bertoglio, Chatenay Malabry, France) diluted 1 : 200 in NaCI/P,/l% gelatin. Rho/GDI and Rho proteins were detected after incubation with peroxidase-coupled anti-mouse IgG by peroxidase staining and chemiluminescence, respectively.

RESULTS Increase in the C3-catalyzed ADP-ribosylation of Rho proteins in differentiated F9 cells

In order to study the involvement of Rho proteins in cellular differentiation, we used the specific ADP-ribosylation by C3 as an indication for regulatory changes of Rho. As shown in Fig. lA , no differences in ["PIADP-ribose incor- poration were observed between H4IIE and H5 cells (Fig. 1 A). In contrast, total cell extracts from differentiated F9 cells revealed an about sixfold increase in the C3-mediated ADP-ribosylation of Rho as compared with non-differentiated F9 cells (Fig. 1 A). After separation into crude cytosolic and membrane fractions, differentiated F9 cells showed approximately threefold and sixfold elevated levels of ADP-ribosylated proteins in cytosolic and membrane fractions, respectively. Identical results were obtained after frac- tionation of total cell extract by ultracentrifugation (45 min,

Because the extent of the ADP-ribosylation of Rho by C3 is affected by the guanine nucleotide bound [5, 14, 33, 341, we next studied the influence of GDP on Rho ADP- ribosylation in differentiated and non-differentiated cells. Thereby we wanted to analyse whether differences in guanine nucleotide binding might interfere with the observed stimulated Rho ADP-ribosylation in differentiated F9 cells. As shown in Fig. 1 B, after GDP-loading 32P-labeling of Rho proteins increased in membrane fractions from non-differentiated cells (F9 Con). In membrane fractions from differentiated F9 cells (F9 Diff) no change in the ADP-ribosylation of Rho was observed after GDP loading (Fig. 1B). Al- though GDP loading caused a large reduction in the ADP- ribosylation of Rho in the cytosolic fraction, no major differentiation-dependent effects of GDP or GTP[S] were observed in cytosol (not shown). The observation that GDP exchange failed to induce changes in the ADP-ribosylation of membranous Rho from differentiated cells might be explained by the assumption that GDP is already bound. If this hypothesis is true, then GTP[S] exchange should cause a reduction in the ADP-ribosylation of membranous Rho from differentiated cells. As shown in Fig. lC , this appears to be the case. GTP[S] decreased the ADP-ribosylation in membranes from differentiated cells by about 60%. In non-differentiated cells the GTP[S] effect was up to 10% (not shown). As a control for complete nucleotide exchange, experiments were performed with recombinant GST-RhoA and recombinant RhoA in the presence of [3H]GTP. These experiments showed > 90% efficiency in nucleotide exchange.

In order to analyse whether the ADP-ribosylation of only one or all Rho species (RhoA, RhoB and RhoC) was en-

1oooooxg, 4°C).

Rho

B

Rho

C fi cn5r' O ~ C o a a

Rho

Fig. 1. ADP-ribosylation of total cell extracts (A) and membrane fractions (B, C) from differentiated (H4IIE, F9 Diff) and non- differentiated (HS, F9 Con) cells. Differentiation of F9 cells was induced by retinoic acid and (Bt),cAMP treatment in low-serum medium as described [40]. Preparation of cell extracts and ['ZP]ADP-ribosylation by C3 were performed as described in Mate- rials and Methods. After SDS gel electrophoresis, gels were dried and Rho proteins detected by autoradiography. (A) 20 pg protein from total extracts was used for ADP-ribosylation. (B) Prior to ADP-ribosylation, membrane fractions were isolated by centrifugation (20min, IOOOOXg, 4°C) and GDP exchange reactions were performed in the presence of 300 pM GDP for 10 min at 30°C as described in Materials and Methods. Con, ADP-ribosylation without GDP exchange; + GDP, ADP-ribosylation after GDP exchange. (C) Membrane fractions from differentiated F9 cells were ADP-ribosylated without nucleotide exchange (Con), after GDP exchange (GDP) and after GTP[S] exchange (GTP[S]). Exchange reactions (300 pM guanine nucleotide) were performed according to B.

hanced upon differentiation, two-dimensional gel electrophoresis was performed. As shown in Fig. 2, the major ADP- ribosylated Rho protein in both differentiated and non-differentiated F9 cells is RhoA. This was concluded from the comparison with two-dimensional gels of [12P]ADP-ribosylated platelet membranes, which contain exclusively RhoA [48] and from the expected isoelectric point which was calcu- lated from cDNA analysis 1281. ADP-ribosylated RhoC and RhoB proteins were only poorly detectable. Thus, the increase in the ADP-ribosylation after differentiation is mainly due to an increased ADP-ribosylation of RhoA.

It has been reported that the extent of the C3-mediated ADP-ribosylation of Rho can also be influenced by regulatory Rho-binding factors like the guanine nucleotide dissociation-inhibiting factor Rho-GDI, which is localized in cytosol [23, 35, 361 and binds to Rho in 1 : l ratio [23]. Low concentrations of detergent (SDS) are reported to release Rho from Rho-GDI, thereby largely enhancing its ADP-ribosyla-

912

+ -H A

Fig.2. RhoA is the major ADP-ribosylated Rho protein in both differentiated and non-differentiated F9 cells. Total extract from non-differentiated (30 pg protein) and differentiated (10 pg protein) F9 cells was ADP-ribosylated as described in Materials and Meth- ods. Subsequently, proteins were separated by two-dimensional gel electrophoresis ; the autoradiography is shown. - H i , indicates the direction of the H ' gradient; (1) RhoA; (2) RhoC; (3) RhoB.

A F9 I Con Diff I

Rho

SDS - + - + (0.01 %)

B F9 ' c I

o n 0 s

Rho-GDI

Fig. 3. Interference of Rho-GDI with the ADP-ribosylation of Rho in differentiatedhon-differentiated F9 cells. (A) ["PIADP- ribosylation of 20 pg protein from cytosolic fractions from differentiated (Diff) and non-differentiated (Con) F9 cells in the absence (-) or presence (+) of detergent (0.01% SDS). (B) 50 pg protein from total cell extracts separated by SDS gel electrophoresis and blotted on nitrocellulose. Rho-GDI proteins were detected by hybridization with a monoclonal anti-Rho-GDI antibody (JK-5 [46] ; diluted 1 : 200 in NaCl/P,/l% gelatin) and subsequent peroxidase staining.

tion by C3 [35, 361. Thus, if less complexation of Rho with GDI causes the differentiation-dependent stimulation of ADP-ribosylation, SDS should increase the [32P]ADP-ribosylation of Rho in non-differentiated cells to the level of differentiated cells but should not affect the ADP-ribosylation of Rho from differentiated F9 cells. As shown in Fig. 3A, SDS enhanced the [-"P]ADP-ribosylation of cytosolic Rho proteins jn extracts from non-differentiated F9 cells. How-

c 0 % B o n

46 kDa - GST-RhOA Rho

Rho 24 kDa -

Fig. 4. Rho proteins are not differentiation-specifically com- plexed. (A) 20 pg total protein extracts from differentiated (F9 Diff) and non-differentiated F9 cells (F9 Con) were incubated in the presence of 300 ng recombinant GST-RhoA fusion protein for 30 min on ice. Subsequently ['ZP]ADP-ribosylation was performed and the reaction products separated by SDS gel electrophoresis. Endogenous Rho proteins and recombinant GST-RhoA fusion protein were detected by autoradiography. (B) 20 pg [32P]ADP-ribosylated proteins from non-differentiated (F9 Con) and differentiated (F9 Dim F9 cells were separated on a non-denaturing, 5% acrylamide gel; the autoradiography is shown.

ever, ADP-ribosylation was stimulated also in differentiated cells. Furthermore, the amount of Rho-GDI protein detected by immunoblotting with anti-GDI antibody was independent of the differentiation state (Fig. 3B). Thus, variations in the association of Rho with Rho-GDI did not cause the differentiation-dependent stimulation of the ADP-ribosylation of cytosolic Rho proteins.

To study whether other factors might stimulate the ADP- ribosylation of Rho in differentiated cells by complexing with Rho, we preincubated the recombinant GST-RhoA fusion protein with cell extracts prior to ADP-ribosylation. The fusion protein can easily be distinguished from endogenous Rho proteins by its higher molecular mass. The ADP-ribosylation of GST-RhoA was largely decreased in the presence of cell extract from both differentiated and non-differentiated cells (Fig. 4A). However, no differences in the ADP-ribosylation of GST-RhoA were observed after incubation with extracts from differentiated or non-differentiated cells. Further- more, separation of [3'P]ADP-rib~~ylated extracts on non- denaturing gels failed to detect additional ADP-ribosylated complexes after differentiation (Fig. 4B). Thus, it seems un- likely that the increased ADP-ribosylation of Rho is due to changes in the association of Rho with other factors.

To exclude that an increased de-novo synthesis of Rho was responsible for the increase in ADP-ribosylation of Rho proteins from differentiated F9 cells, we measured the amounts of rho mRNA and of Rho protein. As shown in Fig. 5A, rho expression (as related to expression of glyceraldehyde-3-phosphate dehydrogenase) was not enhanced in differentiated F9 cells. Furthermore, the amounts of immuno- logically detectable Rho protein was similar in differentiated and non-differentiated F9 cells (Fig. 5B). These data suggest that the differentiation-dependent stimulated ADP-ribosylation of Rho is mainly caused by post-translational mechanisms.

Protein-biosynthesis-dependent increase in the C3-catalyzed ADP-ribosylation of Rho by CAMP and retinoic acid

In combination with low serum concentrations, retinoic acid and (Bt),cAMP are both necessary for a complete phe-

913

Rho

GAPDH Rho

Rho +

Fig. 5. rho mRNA and Rho protein amounts are not elevated in differentiated F9 cells. (A) 20 pg total RNA from non-differentiated (F9 Con) and differentiated F9 (F9 Diff) cells were separated on 1% agarose gels. After transfer onto nylon membrane (Hybond N+), rho mRNA was detected by hybridization with a rhoA cDNA probe as described in Materials and Methods. As internal reference, filters were rehybridized with a glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA probe. (B) 50 pg protein from total extracts was separated by SDSPAGE. After semi-dry blotting, RhoA proteins were detected with a monoclonal anti-Rho antibody [47] by chemiluminescence. The arrow indicates RhoA.

notypical differentiation of F9 cells [38, 391. Next we tested whether phenotypic differentiation is really necessary to affect ADP-ribosylation or whether retinoic acid or (Bt),cAMP alone are sufficient as inducers. As shown in Fig. 6, treatment of F9 cells with (Bt),cAMP or retinoic acid for 24 h increased ADP-ribosylation of Rho by about 5-fold and 2.5- fold, respectively. In contrast, in NIH3T3 cells (Bt),cAMP failed to induce the same effects on Rho ADP-ribosylation as detected in F9 cells (Fig. 6A). Furthermore, the increase in ADP-ribosylation of Rho in F9 cells was time-dependent. A significant (> 50%) increase in ADP-ribosylation occurred 8 h after (Bt),cAMP addition (Fig. 6B). Although the ADP- ribosylation of Rho was largely changed upon (Bt),cAMP or retinoic acid treatment, no change in morphology of the treated F9 cells was observed (not shown). The delayed (Bt),cAMP response indicates that activation of protein ki- nase A did not directly stimulate Rho ADP-ribosylation, but that more complex signal pathways had to be induced. Inter- estingly, (Bt),cAMP had to be present over the whole incubation period to increase Rho ADP-ribosylation. If the medium was changed 4 h after (Bt),cAMP addition and cells were further incubated for 20 h in the absence of (Bt),cAMP, no increase in the ADP-ribosylation of Rho was detected (not shown) .

Inhibition of protein biosynthesis by cycloheximide pre- vented the increase in ADP-ribosylation of Rho induced by (Bt),cAMP and retinoic acid (Fig. 7A); rho mRNA expression was not enhanced upon drug treatment (Fig. 7B). Be- cause changes in the association of Rho with Rho-GDI or other factors have not been observed (see Figs 3 and 4), we studied whether post-translational modifications like phos-

5 600 la .- c - $ - e ? 200

n a a

4 ecAMP

8 16 24

Time (h) Fig. 6. Inducibility of Rho ADP-ribosylation by retinoic acid and (Bt),cAMP in non-differentiated F9 cells. (A) Logarithmically growing F9 or NIH3T3 cells were treated with 0.2 pM retinoic acid (RA) or 1 mM (Bt),cAMP (CAMP) for 24 h. [32P]ADP-ribosylation was performed with 20 pg protein from total cell extracts. Reaction products were separated by SDS gel electrophoresis and ADP-ribosylated products detected by autoradiography. Con = untreated control. (B) Time-dependent increase in ['*P]ADP-ribosylation of Rho after cAMP treatment of F9 cells. 4-24 h after cAMP addition (1 mM) cells were harvested and 20 pg protein from total extracts used for [3*P]ADP-ribosylation. Afterwards, reaction products were separated by SDS gel electrophoresis and the autoradiography quan- titatively evaluated by density analysis. The data shown represent the mean of at least two independent experiments. ADP-ribosylation of total extracts from untreated control cells was set to 100%.

phorylation were responsible for the differences in ADP-ribosylation. Prior treatment of cytosolic fractions from differentiated F9 cells with protein phosphatase type-1 caused a decrease in ADP-ribosylation to the level of non-differentiated F9 cells (Fig. 8), whereas such treatment of control cytosol from non-differentiated F9 cells resulted only in minor decrease in ADP-ribosylation. The ADP-ribosylation of membrane-associated Rho proteins was not influenced by this treatment (not shown).

Influence of proliferation on the ADP-ribosylation of Rho In general, differentiation processes are accompanied by

a decrease in proliferation rate. Thus, we studied whether changes in proliferation might trigger the described increase in Rho ADP-ribosylation which occurred concomitantly with differentiation. These experiments were performed with NIH3T3 cells because, in contrast to F9 cells, fibroblasts are easily arrested in cell cycle by serum deprivation. Further- more, 3T3 cells are preferred for such experiments because the expression of the proto-oncogene c-fos can be monitored after a subsequent serum stimulation as internal control for a successful arrest of growth. The amounts of c-fos mRNA have been shown to increase transiently after serum stimulation of growth-arrested cells or in the presence of cycloheximide [49, 501. As shown in Fig. 9 (A, B), neither inhibition of protein biosynthesis by cycloheximide nor serum sti-

914

A F9 Con Diff I

I cAMP R A '

- - - - CAMP - - + + RA cx - + - + - + - +

+ + - - - - - -

B F9

1 RA I CAMP'

GAPDH

Fig. 7. The stimulation of the C3-catalyzed ADP-ribosylation of Rho after (Bt),cAMP (CAMP) and retinoic acid (RA) treatment of F9 cells depends on protein biosynthesis. (A) 24 h after seeding, F9 cells were treated with 1 mM (Bt),cAMP for 14 h (CAMP) or 0.2 pM retinoic acid for 24 h (RA). Cycloheximide (Cx, 5 pg/ml) was added together with cAMP or 10 h after RA addition, to the medium. ['2P]ADP-ribosylation by C3 was performed with 20 pg protein from total cell extracts. The autoradiography after SDS gel electrophoresis is shown. (B) Total RNA was prepared 0-24 h after cAMP or RA treatment. For Northern analysis 20 pg total RNA was separated on a 1 % agarose gel and subsequently transferred onto a Hybond N' membrane. Hybridization with the vhoA and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA probes were performed as described in Materials and Methods.

mulation of serum-arrested cells influenced the level of the [12P]ADP-ribosylation of Rho by C3. However, both treatments resulted in the expected transient increase in c-fos mRNA (not shown). Deduced from these data, with the pre- caution that different cells were studied, it seems likely that changes in the proliferation rate during differentiation of F9 cells did not interfere with the ADP-ribosylation of Rho. Growth arrest by confluence caused an increase in the ADP- ribosylation of Rho which decreases rapidly again after re- seeding (Fig. 9C). However, because Rho proteins are also thought to be involved in cellular adhesion [51], cell-cell contact or spreading effects might superimpose proliferation effects on Rho in confluent or reseeded NIH3T3 cells.

DISCUSSION

Rho proteins appear to be involved in the regulation of the actin cytoskeleton [ l l , 26, 27, 491. Here, we investigated whether Rho proteins interfere with cellular differentiation. We took advantage of the fact that Rho proteins are specifically ADP-ribosy lated by the clostridial exoenzyme C3. Rho proteins (e.g. RhoA, RhoB and RhoC) are described as specific in vi tm substrates for C3 [I -3, 9-17]. Other members of the Rho family like Rac proteins are only very poor in vitro substrates for C3 152, 541 as compared with Rho. The same is true for RhoG (unpublished results) which has recently been described by Vincent et al. [53]. Changes in the

F9 Con F9 Diff Fig. 8. Phosphorylations sensitive to protein phosphatase type 1 are responsible for the increased ADP-ribosylation of cytosolic Rho proteins from differentiated F9 cells. 10 Fg cytosolic extract from non-differentiated (F9 Con) and differentiated (F9 Dim F9 cells were dephosphorylated for 2 h by protein phosphytase type-1 (PPl) as described in Materials and Methods. Subsequently ADP- ribosylation was performed. Reaction products were separated by SDS gel electrophoresis and [32P]ADP-ribosylated proteins detected by autoradiography. The quantitative density analyis represents val- ues from three independent experiments. The data refer to the ADP- ribosylation of cytotolic Rho proteins from non-differentiated F9 cells which was set to 1.0. +PP1 and -PP1, ADP-ribosylation with (+) or without (-) protein phosphatase type-1 pretreatment. The insert shows the autoradiography of one representative experiment.

C3-induced ADP-ribosylation can be assumed to reflect regulatory changes on the GTP-binding protein Rho. After differentiation of F9 teratocarcinoma cells to neuronal-like cells [41], we observed an about sixfold increase in the ADP- ribosylation of Rho proteins by C3. As determined by two- dimensional gel electrophoresis, the increase in the C3-mediated ADP-ribosylation specific for RhoA but not for other Rho isoforms. The differentiation-induced increase in Rho ADP-ribosylation was detected in both cytosolic and membrane fractions, but appeared to be higher in the membranes. Because differentiated F9 cells showed no increase in rho mRNA nor in the amount of RhoA protein compared to non- differentiated F9 cells, it seems likely that post-translational modifications of Rho or of Rho-associated factors are responsible for this effect.

It has been reported that the ADP-ribosylation of Rho is influenced by the guanine nucleotides bound [2, 331 or by Rho-associated factors like Rho-GDI [35, 361. Thus, we tested whether these parameters changed upon differentiation. The ADP-ribosylation of membrane-associated Rho proteins was stimulated by GDP loading in non-differentiated F9 cells. The same effect was observed in membranes from H41IE and H5 cells (not shown). However, GDP failed to stimulate the ADP-ribosylation of Rho proteins located in membrane fractions from differentiated F9 cells. Interest- ingly, in membranes from non-differentiated F9 cells, GDP caused an increase in the ADP-ribosylation of Rho up to the level of differentiated F9 cells. One possible explanation for this observation is that membrane-associated Rho from differentiated F9 cells has predominantly GDP bound. This hypothesis is consistent with the observation that GTP[S] ex-

915

Cycloheximide FCS-Stimulation I after seeding

Rho Rho Rho

Fig. 9. Influence of proliferation on the C3-catalyzed ADP-ribosylation of Rho proteins in NIH3T3 cells. (A) Logarithmically growing NIH3T3 cells were treated with 10 p g / d cycloheximide. Up to 4 h later, cells were harvested and ['*P]ADP-ribosylation by C3 analysed with 10 pg protein from total cell extracts. (B) Logarithmically growing cells were growth arrested by serum deprivation for 48 h with only 0.5% fetal calf serum (FCS) and then refed with medium containing 20% FCS. Total cell protein was prepared up to 4 h after serum addition as described in Materials and Methods. [32P]ADP-ribosylated proteins were separated by SDS gel electrophoresis and detected by autoradiography. (C) Confluent, serum-deprived NIH3T3 cells were reseeded in medium containing 10% FCS. Up to 8 h after seeding cells were harvested and extracts [3ZP]ADP-ribosylated by C3 ; the autoradiography is shown.

change caused a decrease in the ADP-ribosylation of membranous Rho protein from differentiated cells. Thus, we ten- tatively suggest that the exchange of GDP of membranous Rho protein is reduced with differentiation. GDP/GTP-exchange experiments with cytosolic Rho gave no evidence for the involvement of guanine nucleotides in the differentiation- specific increase in the ADP-ribosylation of cytosolic Rho proteins. Thus, regulatory mechanisms other than changes in guanine-nucleotide binding influence cytosolic Rho proteins during differentiation.

To study whether Rho-associated proteins might cause the increase in ADP-ribosylation of cytosolic Rho proteins from differentiated F9 cells, recombinant GST-RhoA fusion protein was tested as a template to identify Rho-binding proteins which stimulate ADP-ribosylation. However, ADP-ribosylation of the GST-RhoA fusion protein exhibited no differences in the presence of extracts from differentiated or non-differentiated cells. Possible interpretations of these findings are (a) such proteins do not exist, (b) they are com- pletely bound to endogenous Rho proteins or (c) their binding to Rho depends on post-translationally modified Rho (not the case for the recombinant GST-RhoA fusion protein). Af- ter separation of [3ZP]ADP-ribosylated extracts from differentiated F9 cells by native gel electrophoresis, no qualita- tive differences (e.g. the appearance of additional complexes) were observed as compared with non-differentiated cells. Again, this finding does not support the hypothesis that Rho complexes differentiation-specifically.

Furthermore, we have analysed the association of Rho with the guanine-dissociation inhibitor, whose binding to Rho requires post-translationally processed Rho [22,54]. Cy- tosolic fractions from both differentiated and non-differentiated F9 cells showed an identical relative (3-4-fold) increase in ADP-ribosylation in the presence of SDS. This detergent is thought to dissociate the cytosolic Rho-GDI complex, thereby restoring the ability of Rho to serve as substrate for C3 [35]. The amount of GDI was not changed with differentiation. Thus, these findings argue against a role of GDI in differentiation-dependent changes of Rho ADP-ribo- s ylation.

The observation that not only the combination but also (Bt),cAMP or retinoic acid alone increased ADP-ribosylation of Rho proteins in non-differentiated F9 cells, indicate a de- fined signal transduction pathway which might be involved in Rho regulation. It is noteworthy that (Bt),cAMP or retinoic acid increased ADP-ribosylation without inducing changes in cell morphology. For phenotypic differentiation, both drugs have to be present simultaneously. These data

indicate that a complete phenotypic differentiation to neuronal-like cells is not necessary, but that the initiation of the differentiation process by retinoic acid or (Bt),cAMP is sufficient to induce regulatory changes of Rho.

Interestingly, the ADP-ribosylation of Rho stimulated by (Bt),cAMP and retinoic acid depended on protein biosynthesis, although Rho amounts are not increased after drug treatment. The stimulating effect of (Bt),cAMP on the ADP-ribosylation of Rho is cell-type-specific because NIH3T3 cells failed to show a similar (Bt),cAMP response as F9 cells. These data suggest that the synthesis of a protein factor specifically causes an increased ADP-ribosylation of Rho after (Bt),cAMP or retinoic acid treatment in F9 cells. Because we obtained no evidence for complexation of an additional protein factor with Rho, we suggest that protein-synthesis- dependent post-translational modification of Rho or of Rho- associated factors is more likely. One important post-translational modification which recently has been described to interfere with the C3-catalyzed ADP-ribosylation of Rho is phosphorylation [55]. Indeed, dephosphorylation of cytosolic extracts from differentiated F9 cells with protein phosphatase type-1 reduced the subsequent ADP-ribosylation of Rho to the level of non-differentiated F9 cells. These data suggest that phosphorylation(s) events sensitive to protein phosphatase type- 1 occur during differentiation.

It is generally accepted that differentiation processes are paralleled by decreased proliferation. Because F9 cells can- not be arrested in cell cycle by serum depletion, the basic effects of proliferation on the ADP-ribosylation of Rho were tested in NIH3T3 fibroblasts. Serum induction of growth- arrested NIH3T3 cells had no influence on the ADP-ribosylation of Rho. Although these experiments were performed with NIH3T3 cells and have to be interpreted with caution, the studies indicate that proliferation per se does not influence Rho ADP-ribosylation. This hypothesis is consistent with the finding that a complete differentiation of F9 cells (= maximal decrease in proliferation) is not required to increase the ADP-ribosylation of Rho. Treatment of logarithmically growing F9 cells with (Bt),cAMP for 8 h already induced changes in ADP-ribosylation of Rho. Thus, it appears that the increased ADP-ribosylation of Rho in differentiated F9 cells (as compared to non-differentiated F9 cells) reflects changes in Rho regulation which might be due to a tissue-specific (neuronal-like) gene expression after (Bt),cAMP/retinoic-acid-induced differentiation. This hypothesis is in agreement with the recent observation that the ADP-ribosylation of Rho by C3 shows tissue specificity [56, 571, indicating tissue-specific variations in the regulation and

916

physiological activity of Rho proteins. Furthermore, H4IIE and H5 cells, which are both rat hepatoma cells and are distinguished only by their differentiation status, do not differ in the ADP-ribosylation of Rho.

This study was supported by the Deutsche Forschungsge- meinschufi, Sonde~ur~schungsbereich 246. We would like to thank Dr Y. Takai (Kobe, Japan) for providing the Rho-GDI antibodies, Dr J. Bertoglio (Chatenay Malabry, France) for providing RhoA antibody and Dr S. Klumpp (Tubingen, Germany) for protein phosphatase type-I .

REFERENCES 1. Aktories, K. , Weller, U . & Chhatwal, G. S. (1987) FEBS Lett.

2. Aktories, K.. Rosener, S., Blaschke, U. & Chhatwal, G. S.

3. Rubin, E. J., Gill, M., Boquet, P. & Popoff, M. P. (1988) Mol. Cell. Biol. 8. 418-426.

4. Just, I., Mohr, C.. Schallehn, G., Menard, L., Didsbury, J. R., Vdndekerckhove, J., Van Damme, J. & Aktories, K. (1992) J. Biol. Clzem. 267, 10274-10280.

5. Just, I., Schallehn, G. & Aktories, K. (1992) Riochem. Biophys. Res. Cornmun. 183, 931 -936.

6. inoue, S., Sugai, M.. Murooka, Y., Paik, S.-Y., Hong, Y. M., Ohgai, H. & Suginaka. H. (1991) Biochem. Biophys. Res. Commun. 174, 459-464.

7. Sugai, M., Hashimoto, K., Kikuchi, A., Inoue, S., Okumurd, H., Matsumota, K., Goto, Y., Ohgai, H., Moriishi, K., Syuto, B., Yoshikawa, K., Suginaka, H. & Takai, Y. (1992) J. Biol. Chem. 267, 2600-2604.

8. Sugai, M., Chem, C.-H. & Wu, H. C. (1992) Proc. Nut1 Acad. Sci. USA 89, 8903-8907.

9. Aktories. K.. Mohr. C. & Koch. G. (1992) Curr Too Microhiol.

212, 109-113.

(1988) Eur. J . Biochc~n. 172, 445-450.

12.

13.

14.

15.

16. 17.

18.

19.

20.

21.

22.

Inzmun'ol. 175, 1 is- 13 1 . 10. Aktories, K., Braun, U., Rosener, S., Just, I. & Hall, A. (1989)

11. Chardin. P., Boquet, P., Madaule, P., Popoff, M. R., Rubin, E.

Kikuchi, A,, Yamamoto, K., Fujita, T. & Takai, Y. (1988) J.

Narumiya, S., Sekine, A. & Fujiwara, M. (1988) J. Biol. Chem.

Braun, U., Habermann, B., Just, I., Aktories, K. & Vandekerck-

Biochem. Riophys. Res. Commun. 158, 209-213.

J . & Gill, D. M. (1989) EMBO J. 8, 1087-1092.

Bid. Chen?. 263, 16303-16308,

263, 17255-17257.

hove. J . (1989) FEBS Lett 243, 70-76.

Res. 16, 2117. Chardin, P.. Madaule, P. & Tavitian, A. (1988) Nucleic Acids

Madaule, P. & Axel, R. (1985) Cell 41, 31-40. Yeramian, P., Chardin, P., Madaule, P. & Tavitian, A. (1987)

Nucleic Acids Rex 15, 1869. Bourne, H. R., Sanders. D. A. & McCormick, F. (1991) Nature

349, 117-127. Fukumoto, Y., Kaibuchi, K., Hori, Y., Jujioka, H., Araki, S.,

Ueda, T., Kikuchi, A. & Takai, Y. (1990) Oncogene 5, 1321 - 1328.

Garrett, M. D., Self, A. J., Van Oers, C. & Hall, A. (1989) J.

Kikuchi, A,, Kuroda, S., Saski, T., Kotani, K., Hirata, K., Katay- ama, M. & Takai. Y. (1992) J. Bid. Chem. 267, 14611- 34615.

Mizuno, T., Kaibuchi, K., Yamamoto, T., Kawamura, M., Sa- koda, T., Fujioka. H., Matsuura, Y. & Takai, Y. (1991) Proc. Nut1 Acad. Sci. USA 88. 6442 - 6446.

Biol. Chern. 264, 10-13.

23.

24.

25.

26.

27.

28.

29.

30. 31.

32.

33. 34.

35.

36.

37. 38.

39. 40. 41.

42.

43. 44. 45.

46.

Regazzi, R., Kikuchi, A,, Takai, Y. & Wollheim, C. B. (1992)

Settleman, J., Albright, C. F., Foster, L. C. & Weinberg, R. A.

Sekine, A., Fujiwara, M. & Narumiya, S. (1989) J . Bid. Chenz.

Paterson, H. F., Self, A. J., Garrett, M. D., Just, I., Aktories, K. & Hall, A. (1990) J. Cell Biol. 111, 1001-1007.

Wiegers, W., Just, I., Muller, H., Hellwig, A,, Traub, P. & Akto- ries, K. (1991) Eur. J . Cell Biol. 54, 237-245.

Lang, P., Guizani, L., Vitte-Moni, I., Stancou, R., Dorseuil, O., Gacon, G. & Bertoglio, J. (1992) J. Biol. Chern. 267, 11 677- 11 680.

Stasia, M.-J., Jouan, A., Bourmeyster, N., Boquet, P. & Vignais, P. V. (1991) Biochenz. Biophys. Res. Commun. 180, 615-622.

Hirata, K. & Takai, Y. (1992) J. B i d . Chem. 267, 8719-8722. Avraham, H. & Weinberg, R. A. (1989) Mol. Cell. Bid. 9,

Avraham, H. (1990) Biochem. Biophys. Rex Commun. 168,

Aktories, K. & Frevert, J. (1987) Biochem. J. 247, 363-368. Habermann, B., Mohr, C., Just, I. & Aktories, K. (1991) Bio-

chirn. Biophys. Acta 1077, 253-258. Bourmeyster, N., Stasia, M.-J., Garin, J., Gagnon, J., Boquet,

P. & Vignais, P. V. (1992) Biochemistry 31, 12863-12869. Just, I., Mohr, C., Habermann, C., Koch, G. & Aktories, K.

(1993) Biochem. Phurmacol. 45, 1409 -141 6. Darlington, G. J. (1987) Methods Enzymol. 151, 19-38. Jakob, H. & Nicolas, J.-F. (1987) Methods Enzymol. 151, 66-

Kuff, E. L. & Fewell, J. W. (1980) Dev. Biol. 77, 103-115. StricMand, S. & Mahdavi, V. (1978) Cell 15, 393-403. Chomczynski, P. & Sacchi, N. (1987) Anal. Biochem. 162.

156-159. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular

cloning: a laboratory manual, Cold Spring Harbor Labora- tory Press, Cold Spring Habor NY.

J. Biol. Chem. 267, 17512-17519.

(1992) Nature 359, 153-154.

264, 8602-8605.

2058-2066.

11 4- 124.

85.

Bradford, M. M. (1976) Anal. Biochenz. 72, 248-254. Laemmli, U. K. (1970) Nature 227, 680-685. Harris, E. L. & Angal, S. (1989) Protein pur(fication methods,

a practical approach, p. 29, IRL press, Eynsham. Shimizu, K., Kaibuchi, K., Nonaka, H., Yamamoto, J. & Takdi,

Y. (19911 Biochem. Bioahvs. Res. Conzmun. 175. 199-206. \ I 1 ,

47. Lang, P., Gesbert, F., Thiberge, J. M., Troalen, F., Dutartre, H.. Chavrier, P. & Bertoglio, J. (1993) Biochern. Biophys. Res. Commun. 196, 1522-1528.

48. Nemoto, Y., Namba, T., Teruuchi, T., Ushikubi, F., Morii, N. & Narumiya, S. (1992) J . Bid. Chem. 267, 20916-20920.

49. Buscher, M., Rahmsdorf, J. L., Litfin, M., Karin, M. & Herrlich, P. (1988) Oncogene 3, 301-311.

50. Rahmsdorf, H. J., Schonthal, A,, Angel, P., Litfin, M., Ruther. U. & Herrlich, P. (1987) Nucleic Acids Res. 15, 1643-1649.

51. Ridley, A. J. & Hall, A. (1992) Cell 70, 389-399. 52. Menard, L., Tomhave, E., Casey, P. J., Uhing, R. J., Snyderman.

R. & Didsbury, J. R. (1992) Eur: J. Biochem. 206, 537-546. 53. Vincent, S., Janteur, P. & Fort, P. (1992) Mol. Cell. Biol. 12.

54. Hori, Y., Kikuchi, A,, Isomura, M., Katayama, M., Miura, Y., Fujioka, H., Kaibuchi, K. & Takai, Y. (1991) Oncogene 6, 515-522.

3138 - 3148.

55. Fritz, G. & Aktories, K. (1993) Biochem. J. 300, 133-139. 56. Morii, N., Akihiro, S., Ohashi, Y., Nakao, K., imura, H., Fuji-

wara M. & Narumiya, S. (1988) J. Biol. Chem. 263, 12420- 12426.

57. Fritz, G., Lang, P. & Just, I. (1994) Biochim. Biophys. Acta, in the press.

differentiation-induced increase in clostridium botulinum c3 exoenzyme-catalyzed adp-ribosylation of...

Documents